Empirical relationships between the macroseismic intensity and the ground‐motion parameters for the Himalayan region are derived in this study. A strong‐motion database from 21 moderate‐to‐large earthquakes, along with their corresponding macroseismic intensity, is considered. All the intensity values are inferred from isoseismal maps and earthquake damage reports and then converted to the modified Mercalli intensity (MMI) scale. An orthogonal regression analysis is used to find the best correlation between MMI and peak ground acceleration (PGA), peak ground velocity (PGV), and pseudospectral acceleration (PSA) at 0.3, 1.0, 2.0, and 3.0 s to accommodate the uncertainty in the regression coefficients. In addition to the ground‐motion parameters, the MMI is related to other potential independent variables, including the moment magnitude, the hypocentral distance, and the 30‐m average shear‐wave velocity (VS30). The study shows that site effect is noticed predominantly in the MMI–PGA relationship for a hypocentral distance of more than 200 km. When relating MMI with PGV, PSA0.3 s, PSA1.0 s, PSA2.0 s, and PSA3.0 s, however, the contribution of site effects to the correlation is negligible. To eradicate the site effect in the PGA–MMI relationship, the derived empirical relationship is modified, based on the statistical analysis of MMI observed and MMI predicted, with or without including VS30 as a potential independent variable. Furthermore, the developed regression models are verified using several statistical tests, the F‐test, the t‐test, the Durbin–Watson test, and the Breusch–Pagan test. Additionally, the Euclidean distance concept is evaluated in the study. It was concluded that the PGA is a good indicator for deriving the MMI value in the Himalayan region, but that one should use site‐specific MMI versus PGV and MMI versus PSA relationships to predict reliable parameters.